JPS60149260A - Binary coding circuit of picture information - Google Patents

Abstract

PURPOSE:To suppress the increase of information in quantity at the character and linear pattern parts by deleting microspots of low and high densities which will be a cause of the increase of information at the character and linear pattern parts. CONSTITUTION:The outputs of line memories 4a-4n are applied to a white picture element detecting circuit 14 and a black picture element detecting circuit 15. Then perfect white and black picture element detecting circuit 15. Then perfect white and black picture elements are selected for every (m) units of picture elements in the main scan direction and every (n) trains of picture elements in the secondary scan direction. A counter 16 detects the number of selected picture elements. When the number of these selected picture elements is larger than the threshold number of units set at a threshold value memory 9', a comparator 17 selects the simple binary coding. Otherwise a dither pattern method is selected. At the same time, the maximum and minimum densities are detected by maximum and minimum value detecting circuits 6 and 7 respectively. Then a subtractor 8, the memory 9' and a comparator 10 decides that the value obtained by subtracting the minimum density value from the maximum density value is larger than the density difference threshold value. This suppresses the increase of information quantity at the character and linear pattern parts and simplifies the binary coding of picture information.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an image information binarization circuit, and is used in document information processing to compress document information of a document in which grayscale images such as photographs and text coexist. This invention relates to an image information binarization circuit suitable for.

[Background of the invention]

In conventional document information processing devices such as facsimiles, it is necessary to divide a document image vertically and horizontally into small areas (hereinafter referred to as pixel units), each of which has a black density that is whiter than a black density threshold. Depending on whether the pixel is black, either white or black is selected, and the document image is reproduced as a set of pixel units determined to be white or black.

This method is called a binarization process because the black density of each pixel is set to either white or black.

Original documents have light black, that is, gray parts, and even gray has dark gray and light gray, so the black density of each pixel is identified in multiple stages between white and black (in the following, multi-value ), the original document image can be better reproduced.

However, for document images consisting only of characters, lines, etc.
Rather than distinguishing between white and black in multiple stages, it is possible to reproduce a document image with better contrast by selecting either white or black. In addition, the amount of information of a document image is expressed as M-N, where N is the number of pixels and M is the number of stages during multilevel processing. Therefore, when M = 2, which is the minimum value of the number of stages, This method requires the least amount of image information during conversion processing, and is convenient for electrically recording document image information or transmitting it in a short time.

On the other hand, in recent years, image reading devices have become more precise, and the area of each pixel has almost reached the same level as halftone dots used in printing. By changing the number of black dots within an area in proportion to the black density, it is becoming possible to express gray in multiple levels.

An example of this method is the dither pattern method.
The figure is an explanatory diagram showing the outline thereof.

FIG. 1 (1) is an example of reading pixel density, and shows an example of density for 16 reading pixel units.

Further, (2) in the figure is an exemplary diagram showing an example of a dither pattern matrix.

Now, let the density of the read pixel unit of m rows and n columns be S m, and the value of m rows and n columns of the dither pattern matrix is,
7, the pixel unit density T after applying the dither pattern method
1 is given by the following equation.

However, T,, n=1 is black, T-,,=O is 2
The document image is reproduced using values, but by using this method,
Visually, it is possible to read gray as well as white and black.

It should be noted that the conventional simple force binarization is shown using the dither pattern method in the simple binarization explanatory diagram of FIG.

(1) in Fig. 2 is a read pixel density side view, and (2)
) is an illustration of the dither pattern method at the time of binarization), and (3) is an exemplary diagram after processing.

That is, conventional simple binarization is equivalent to processing in which each element of the dither pattern matrix is set to a constant value.

Comparing the simple binarization process and the dither pattern method shown above, the simple binarization process is sufficient for characters and line drawings, and the dither pattern method for photos and gray areas is effective. .

However, when the dither pattern method is used for text and line drawings, it has the disadvantage that the outlines are not sharply cut, and a method to improve this problem is to separate the image part of the text and line drawing from the photograph or gray area. There is a method (hereinafter referred to as image area separation processing) that attempts to apply a dither pattern method and simple binarization processing to each of these methods.

As an example, find the difference between the maximum density and minimum density of each pixel in a pixel block consisting of m pixels in the vertical direction and n columns in the horizontal direction, and if this difference is greater than the threshold value, the simple 2
There is a method that uses value conversion processing and uses a dither pattern method for other processing.

The size of a pixel unit is approximately 1259m x 125μm, and if the values of m and n above are each 4, the size of a pixel block will be an area of approximately 0.5 wa × O, 5 days, and the maximum within that area The difference between the density and the minimum density is almost equivalent to determining the density gradient within a minute area, and the contours of characters and lines have large density gradients, so the dither pattern method described above can be applied to characters and lines. It is possible to partially prevent the distortion applied to the figure, and the contour lines become sharper.

An example of a circuit used for such processing is shown in FIG. 3, which is a conventional image binarization circuit diagram.

That is, the photoelectric conversion element 1 reads the black density of each pixel, and A
The /D conversion circuit 2 converts the black density into an integer value, and the value is sent to each line memory 4 through the switch 3.・.

..., 4. After charging with n lives, press switch 5. , . . . , 51, the black density signal of m pixels in each line is detected by the maximum value detection circuit 6 for detecting the maximum density value.
is sent to the minimum value detection circuit 7 for detecting the minimum density value, and the difference between the values obtained, that is, the value obtained by subtracting the minimum density value from the maximum density value, is calculated by a subtractor 8, and the difference value and the threshold value are calculated. The value in the memory 9 is compared with a comparator 10, and if the value is larger than the threshold value, a signal is output to select a uniform value pattern, and if it is smaller than the threshold value, a signal is output to select a dither pattern.

Based on this signal, the dither method/binarization matrix memo 1-11 outputs the dither pattern method or simple binarization matrix value, and the dither method/binarization circuit 12 for the binarization processing circuit outputs the dither pattern method or simple binarization matrix value. treatment, then MH/MR
Information is compressed in the encoding circuit 13, and this image information is stored or transmitted.

However, as a result of various studies using this circuit, it was found that the Image Electronics Engineers of Japan (Facsimile Test Chart) No. 1-
It was found that the amount of information in the character part of an image using L is the same whether the entire image is processed using the dither pattern method or when it is processed using the circuit shown in Figure 3. This is approximately 2.4 times as large as when the image is simply binarized.

That is, Table 1 shows the values.

Table 1 The values in the table are calculated by dividing the number of encoded information bits by the number of pixels,
It is multiplied by 100 and is generally called the encoded information compression rate.

As mentioned above, conventional image area separation processing is used to separate characters.

Although it is possible to improve the contours of line drawings, etc., it has the disadvantage that the amount of information in text and line drawing areas cannot be compressed in the same way as simple binarization.

[Purpose of the invention]

An object of the present invention is to provide an image information binarization circuit that suppresses the increase in the amount of information in text and line drawing parts, which is a drawback of the prior art.

[Summary of the invention]

The configuration of the image information binarization circuit according to the present invention is such that document pixel information, which is an integer value corresponding to the color density of each pixel obtained by subdividing a document including a grayscale image, is expressed as an integer m of 2 or more in the main scanning direction. For each pixel block consisting of n rows of represented m pixel units in the sub-scanning direction, represented by an integer n of 2 or more, a specified number or more of pixel units in that pixel block have a low density side density. A circuit ■ that determines whether the density is below the threshold; a circuit ■ that determines whether the specified number of pixels in a pixel block is equal to or higher than the high-density side density threshold; and a circuit A circuit ■ that determines whether the value obtained by subtracting the value is equal to or greater than the concentration difference threshold, and the circuits I, ■, and ■
The circuit (2) selects the density of each pixel in the pixel block according to the output from the predetermined amount value or the dither pattern threshold value, and then compares it with the selection threshold value and binarizes the density value.

More detailed information is as follows.

That is, the present invention achieves the object by eliminating the low density and high density minute points which are the cause of the increase in the amount of information in the text and line parts in the conventional method using the circuit shown in FIG. It was designed to be achieved.

For example, if a document image is divided into 16 levels and the black density is read at high speed, the minimum level 1 or the maximum level 16 of the 16 levels may be incorrect due to electrical circuit noise. Sometimes it is recognized.

In the circuit shown in Figure 3, simple binarization processing is performed for areas with large differences in shading, but if there are pale gray points in white areas and slightly faint black points in dotted areas, the dither pattern method is used to process them. It remains as information.

Furthermore, if the density gradient is gentle in the contours of characters and line drawings, the dither pattern method is used.

In the portion to which the dither pattern method is applied, the black and white pixel units appear with great continuity after processing, so the amount of information increases significantly with MH/MR encoding. These are the drawbacks of image area separation processing according to the prior art, and the present invention is intended to solve these problems.

[Embodiments of the invention]

An embodiment of the image information binarization circuit according to the present invention will be described with reference to FIG.

That is, FIG. 4 is an image information binarization circuit diagram according to an embodiment of the present invention.

In the figure, parts with the same symbols as those in Figure 3 above are equivalent parts, 9' is a threshold memory, 14 is a white pixel detection circuit, 15 is a black pixel detection circuit, 16 is a counter, 17 is a comparator, and 18 is a It is a logical sum circuit.

The illustrated line memory 4. 4o, the photoelectric conversion element 1 reads the black density of each pixel in the same way as in FIG. It is something that

Further, the threshold memory 9', the white pixel detection circuit 14, the counter 16, and the comparator 17 correspond to a circuit I that determines whether a specified number or more of pixels in a pixel block is less than or equal to the low density side density threshold. Similarly, the threshold memory gt1 black pixel detection circuit 15, counter 16 and comparator 17 are
This circuit corresponds to the circuit (■) that determines whether the specified number or more of pixels in a pixel block is equal to or higher than the high-density side density threshold, and also includes a maximum value detection circuit 6, a minimum value detection circuit 7, a subtracter 8, and a threshold value. The memory 9' and the comparator 10 correspond to a circuit (2) that determines whether the value obtained by subtracting the minimum density value from the maximum density value of each pixel in a pixel block is equal to or greater than the density difference threshold. 9', the OR circuit 18, the dither method/binarization matrix memory 11, and the dither method/binarization circuit 12 are the same as the circuit 1.9' described above. This corresponds to the circuit (2) that binarizes the density of each pixel in the pixel block according to the outputs of II and III by selecting either a constant value or a dither pattern threshold and then comparing it with the selection threshold. .

That is, the difference between this embodiment and the prior art described in FIG. 3 is that in addition to the configuration shown in FIG. 3, the line memory 4. ,...

The white pixel detection circuit 14 and the black pixel detection circuit 15 output the output from 4 for m pixel units in the main scanning direction and n columns in the sub-scanning direction (where m.

n is an integer of 2 or more), that is, complete white and complete black are selected in each pixel unit in the mxn pixel block, the number is detected by the counter 16, and the number is determined by ′
Comparator 1
In step 7, simple binarization is selected, and in other cases, dither pattern method is selected.

Then, the selection result of this dither method/binarization and the selection result of the conventional method shown in FIG. This is how it was done.

Thus, the threshold value used for each determination is output from the threshold value memory 9', and the output of the threshold value memory 9' related to this threshold value setting section is configured so that it can be changed depending on the previous determination result. .

In other words, if the previous judgment result was simple binarization, the threshold for this judgment can be changed to make it easier for simple binarization to be selected, but the output of this threshold memory is always as a constant. However, it is possible to improve the conventional technology, and the selection thereof is arbitrary.

FIG. 5 is a characteristic diagram showing each threshold value of the comparator 10 in the circuit shown in FIG. 3, the presence or absence of deterioration of the photographic partial image (broken line), and the change in the encoded information compression ratio (solid line).

From this result, the minimum threshold value when simply binarizing the portion where the difference between the maximum value and the minimum value of the pixel unit density in the conventional pixel block exceeds the threshold value is 50 within the range that does not exceed the character area MR
At 6 (when the density is set from O to 16), which is the direction in which the encoded information compression rate decreases, the MR encoded information compression rate for the character part at that time is 24%. It is not possible to compress the amount of information.

However, as shown in FIG. 4 according to the above embodiment,
When determining using the comparator 10 and the comparator 17,
As shown in the characteristic diagram of FIG. 6, the threshold value of the comparator 10 is kept at 7 (in the circuit of FIG. 3, the information compression rate is 25.6% as shown in FIG. 5), and the comparison is made. As the threshold value of the sum of the numbers of complete white pixel units and complete black pixel units of the unit 17 increases, the encoded information compression rate becomes smaller and smaller, but the image quality deterioration is not perceptible at all.

For example, set the threshold for the number of completely white and completely black pixels to 1.
Even if the number of images is 0, the image quality deterioration is within the acceptable range, and the information compression rate reaches 16.5%, which is close to 11.0% when all images shown in Table 1 are simply binarized. Become.

As described above, according to the present invention, the number of pixel units composed of n columns of m consecutive pixel units is different from the maximum density and the minimum density for each mx'n pixel block. By ORing the judgment based on the difference in density and the new judgment based on the sum of the number of pixel units of completely white and completely black, the method of selecting the dither pattern method and simple binarization can be used to remove gray areas in photographs, etc. There is no deterioration in the image quality of the included portion, and it is possible to highly compress the amount of encoded information.

Therefore, the image information binarization circuit according to the present invention can be applied to general image input devices, and in particular, to computer input terminals,
Effective for facsimile, etc.

〔Effect of the invention〕

According to the present invention, it is possible to provide an image information binarization circuit that suppresses the increase in the amount of information in text and line drawing parts, which is a drawback of the prior art, and the invention has excellent practical effects. It can be said that.

[Brief explanation of the drawing]

Figure 1 is a schematic explanatory diagram of the dither pattern method, Figure 2 is
An explanatory diagram of simple binarization, FIG. 3 is a circuit diagram for binarizing image information related to the prior art, and FIG. 4 is a circuit diagram for binarizing image information according to an embodiment of the present invention. The figure is a characteristic diagram of the prior art, and FIG. 6 is a characteristic diagram of the one according to the present invention. 1... Photoelectric conversion element, 2... A/D conversion circuit, 3.
...Switch, 4. ,..., 41... line memory, 5Il,... 5A...Switch, 6...Maximum value detection circuit, 7.・
, minimum value detection circuit, 8... subtractor, 9'... threshold memory, 10... comparator, 11... dither method/binarization matrix memory, 12... dither method/binary value conversion circuit, 13... MH/MR, encoding circuit, 14... white pixel detection circuit, 15... black pixel detection circuit, 16... counter, 17... comparator, 18... OR circuit. Figure 1, (1) (2) (3), Figure 2, Figure 3, Floor 4-Figure 5 (illustrated by Koumi) 31h 16 (F!jJJ) Continuation of page 1 0 Inventor Hideyuki Shio Hitachi Ichisaiwaicho 3-chome office

Claims (1)

[Claims] 1. Document pixel information, which is an integer value corresponding to the color density of each pixel obtained by subdividing a document containing a grayscale image, in the main scanning direction.
For each pixel block consisting of n rows in the sub-scanning direction, representing m pixel units represented by an integer m greater than or equal to 2, a specified number or more of pixel units within that pixel block. a circuit that determines whether the number of pixels in the pixel block is equal to or less than the density threshold on the low-density side; A circuit (■) determines whether the value obtained by subtracting the minimum density value from the maximum density value is greater than or equal to the density difference threshold, and the density of each pixel in the pixel block is determined by a fixed amount according to the outputs of the circuits (2), (2) and (2). 1. A circuit for binarizing image information by comparing the selection threshold with the selection threshold after selecting one of the selection threshold and the dither pattern threshold. 2. In the item described in claim 1, each of the low density side density threshold, high density side density threshold, and density difference dark value in the threshold value setting unit related to determination is set as the processing result of the pixel block of preprocessing. An image information binarization circuit configured to change depending on the image information.